Pensions schemes that apply data matching techniques could see “dramatic improvements in the quality of their member data”, and should use the dashboards programme delay to get their data “dashboard ready”, according to new research sponsored by LCP.
The research focused on member data from a large defined benefit pension scheme with around 17,000 non-pensioner members.
Analysis was conducted in partnership with Digidentity, a certified identity service provider, under the UK’s Digital Identity and Attributes Trust Framework and TransUnion, a global information and insights company and one of the UK’s leading credit reference agencies.
Return to sender
TransUnion was provided with selected information that was compared with its databases to check the accuracy of names, addresses and other personal data.
The pensions dashboards project is reliant on accurate name and address data, as these will be two of the three “verified” data fields sent to pension schemes for data matching.
TransUnion was then asked to identify whether member data was 100 per cent accurate – that is, living as stated – or whether there were discrepancies; for instance, someone now at a new address, a member who had changed their name
TransUnion found that member data was accurate for about five out of six active members, but only just over half (58 per cent) of deferred members.
The largest area for discrepancy was address information, in particular where the address held by the DB scheme appeared to be out of date and a new address was available.
Address information was incorrect for a higher percentage of deferred members (16 per cent) compared with active members (5 per cent). However, even one in 20 active members had moved house and not notified the scheme.
The research also found that around one in 10 deferred members had no address match, but could be matched purely on the basis of name and date of birth against a person showing on TransUnion’s records.
Find the member
A large number of smaller discrepancies were discovered on address data held by the scheme, including around 200 members with missing postcodes – which could be added – and a further 200 where postcodes held were not correct.
Given the importance of addresses for data matching with pensions dashboards, even a small degree of data cleansing could greatly reduce the number of partial matches between user data and scheme data, which would require further work to resolve.
LCP partner Sir Steve Webb said: “The process of preparing for pensions dashboards could have a huge spin-off benefit by leading to a step change in the quality of pension scheme data.
“Both DB schemes and [defined contribution] schemes will have data issues, sometimes relating to the lack of regular contact with members, but also because of inaccuracies in data supplied by employers.
“This research shows that even a relatively simple one-off exercise can lead to a major improvement in scheme data, with a wide range of benefits for pension schemes and providers as well as members.
